Karen Raugust reports how extensive previs created a problem solving road map for British Telecoms acclaimed Networking spot.
If you have the QuickTime plug-in, you can view the completed BT-ICT Networking spot and earlier versions by simply clicking the link.
Previsualization, while rare for commercials, can go a long way toward smoothing the production process for a complex spot. A case in point: BT-ICT Networking, a 60-second TV ad for British Telecom thats been nominated by the VES for Outstanding Visual Effects in a Commercial. It used acrobats portraying businesspeople, architects, doctors and customers to demonstrate BTs network commerce capabilities in a wide range of industries, including commodities, currency, medical and government. The previs process accounted for nearly six weeks of the five months spent creating the spot, compared to the two days to two weeks typically set aside for previs on commercials.
BT-Networking, one component of a £26 million BT advertising campaign, was directed by Joseph Kahn out of London production company Exposure Films through London agency St. Lukes, and executive produced by Kenny Solomon of The Syndicate (TS). TS completed the visual effects at its Santa Monica facility and its CaféFX division in Santa Maria. The spot debuted in September 2004.
The fast-paced action moves through a detailed, fully CG city that combines elements from London, Vancouver, Los Angeles and Ghent. It included live acrobats and fully CG people and animals; many scenes feature seamless transitions from live-action to CG characters and back again. Two to three of the CG people appear in close-ups while the rest are mid-distance or transitional, according to TSs David Lombardi, the digital effects supervisor, who worked closely with visual effects supervisor Eric Durst during previs. TS used LightWave 3D to model and animate the CG city and objects, Maya for character animation and Digital Fusion and Discreet flame for compositing.
The strong treatment and storyboards made it clear there was twice as much information as could easily fit into a 60-second spot. We knew it was a really tight window to get all the ideas across, says Lombardi. We knew whittling down and picking out places to cut would take time. We knew there would be a lot of creative decisions. In addition, everything going on in the spot would require complicated camera moves and split-second timing to capture the acrobatic interaction.
All of that meant TS needed to build in adequate time for planning: Hence, the unusually long six-week previs period. The objective was not only to figure out how to shoot the commercial in terms of timing, direction and equipment needs, but to create a roadmap to help explain all the moves and camera angles to the crew. If you say, `Its a fully greenscreen environment with people flying around, it can get a little ambiguous, comments Lombardi. The work done during previs would help the crew interpret how to make the spot happen.
The previs probably accounted for 5% to 10% of the spots budget, Lombardi estimates. The shoot itself took one to one-and-a-half weeks, with the remaining 12 to 13 weeks spent on post. While 10 to 20 people were working on post duties at a given time, just two primary people were involved during previs.
Previsualization began with the creation of a low-res, geometric 3D animatic using gray block forms to represent the acrobats and other objects. The animatic was matched to the framing stills in the original storyboard. When completed, the first animatic timed out at 1 minute and 50 seconds, nearly double the 56 seconds needed. (The spot ended with a four-second tag.) TS addressed the additional length by first creating a faster pace; the original concept called for high energy. It also tightened up some of the shots. This process eliminated 35 seconds, leaving a still-too-long 1 minute 15 seconds. The next step was to strategically delete some shots and trim others, then reanimate. This version met the 56-second goal.
The process of developing the animatic was completed in three to four weeks. After that, TS spent another two weeks creating a second, greenscreen version, exactly the same as the original animatic, but showing the soundstage, where camera shots and moves could be plotted on measurement grids, which replaced the city background featured in the original version. TS used the soundstage animatic to go back and deconstruct all the shots, selecting layout, focal distances and camera lenses. That was the hard part, coordinating all these elaborate moves, Lombardi says.
TS essentially set up little speedometers on all cameras and people to measure how long it would take to perform each move and how fast people, cameras and objects would have to travel. That dictated whether certain scenes would be possible; for example, a shot that required a camera to speed along a track at 40 miles an hour would not be feasible as originally planned. In one scene, a camera circled 360 degrees around a sequence in which two businessmen on wire rigs manipulated by grips to enable live people to perform the entire move flip over two security guards standing on elevated greenscreen platforms traveling on a track. The 2.5-second scene, which represented a gold/euro currency exchange, couldnt be completed as intended because the camera wasnt able to move fast enough to complete the circle around the action, and because the track would be visible. The shot was altered so the camera, suspended on a rail, moved through the action instead of around it.
TS added an element of flexibility by taking one or two days to make the CG city modular so it could be changed later. For example, buildings could be raised or lowered to alter the amount of sky. While this additional work didnt turn out to be necessary everything fell into place and the client liked the spot TS thought the extra effort was worthwhile since it would have helped solve unanticipated problems that might well have popped up.
Since there were so many decisions to be made in terms of trimming and tightening, TS, Kahn and the client worked together closely during previs. Certain things we leaned toward removing or trimming right away, Lombardi offers. We knew some things were financially or physically impossible. Thus much of the trimming was done during the creation of the original animatic, with further tweaking as the soundstage version was completed.
From the beginning, we were always keeping in mind, `How are we going to shoot this?, adds Lombardi, who explains that problem scenes were trimmed or altered as early in the process as possible. We didnt want the client to get too attached to a shot that wouldnt be possible in a high degree of quality.
Four to five shots had to be eliminated, significantly trimmed or reduced from a featured to a background shot. One scene of a contractor catching a model of a house, for example, was intended as a featured scene in which the camera locked on the rotating model, but it was problematic to shoot and consequently made shorter and less prominent.
On the other hand, certain scenes, such as the money exchange, were too important to eliminate or move to the background. In these cases, TS and Kahn modified camera angles or movements and combined more scenes into one camera shot (in one case integrating four scenes into a single camera angle). They also replaced some real actors with CG versions, such as when there was too much distance or not enough time for a live acrobat to perform a move or hit a mark. Thirteen actors who were in or near the foreground, and therefore had to be more detailed than other all-CG characters, were cyber-scanned for digital conversions.
When a scene was impossible or impractical, TS first went to Kahn. If the director and vfx house were unable to come up with a reasonable solution, they met with the client to explain the creative and technical reasons why an element couldnt be completed as planned, and to present several potential solutions from which to choose. The client understood the need to change if there was a technical limitation, says Lombardi. They knew what they wanted, but it was collaborative.
The shoot itself went smoothly, largely due to the information provided by the animatic, which was followed exactly; Lombardi and editor David Blackburn were able to drop finished scenes right into it. The currency exchange depiction, which, like other scenes, had to be completed with specific choreographed beats, took only about 40 takes a very short time for such a complex concept. Twenty of those took place as the crew worked out how to accomplish the sequence and the other 20 as inevitable mishaps occurred, such as clothing snagging on wire rigs. Lombardi took a small workstation to the location so he could use the real-time soundstage version of the animatic to show people how to shoot some of the wilder camera moves that were easier to convey graphically than verbally.
Lombardi believes more commercial directors should consider previs. A lot of directors have very aggressive concepts, he says. But they dont take advantage of this opportunity. Previs allows directors and vfx studios to nail down how to shoot sequences and better convey that information to the client and crew. When questions come up, there is time to think about the best solution rather than dealing with the issue on the fly, as happens when those same questions arise during the shoot. With previs, the shoot always goes smoothly and no ones ever disappointed, Lombardi suggests. I hope more clients and agencies take advantage of it.
Karen Raugust is a Minneapolis-based freelance business writer specializing in animation, publishing, licensing and art. She is the author of The Licensing Business Handbook (EPM Communications).